PLKNPLC
Optimization of selected proteins
In addition to finding the enzymes with the most endogenous decolourizing ability for MB, we also wanted to optimise the activities of the enzymes which showed the highest percentage of decolourization.
The structure for BBa_K4936001, a Trametes versicolor lignin peroxidase isozyme, was predicted using Alphafold. Chiron, a server provided by Dokholyan Lab at the Pennsylvania State University, was used for further energy minimization of the structure. Crystal structures for our other proteins were retrieved from the PDB database.
We then applied the Alphafill algorithm to reconstitute the protein cofactors with homology based prediction. We then cleaned the structures with pdb-tools, a pypi package publicly available on Github by bonvin lab [1]
Due to time constraints, we were unable to construct and introduce our protein mutants to a biological chassis to analyze their activities in vivo. The in silico tests we were able to perform are detailed below.
Protein docking for all protein structures was performed by SwissDock [2] with accurate presets and 3Å sidechain binding pocket flexibility. Further analysis was performed by UCSF Chimera [3] and visualised in PyMOL. Preliminary protein-ligand interaction analysis was performed in Discovery Studio and ligand coordinates were then extracted from the protein-ligand complex for Molecular Dynamics Simulation in GROMACS. Note that hydrogen atoms were added in OpenBaBel [4], prior to parameterization and topology generation in CGenFF [5].
We have gained a thorough understanding of the optimal methods, parameters, and tools for structure optimization and analysis. To expand on our newfound knowledge and generated data, we plan to do further analysis on our structures to approach the issue of methylene blue removal from different angles. For example, we originally planned to further identify the best proteins of choice with umbrella sampling to test the activity of less likely conformations, and visualising and characterizing their ability on the molecular level with the QMMM/MD functions of the CP2K GROMACS software package.
Our work could also be expanded on by trying to engineer new proteins based on our optimisation data and structures. One possible path is by using flex_DDG [6] to generate a wide sample of mutated structures and model their changes in binding free energies.
We also plan to validate the efficiency of different media in supporting the growth of S. elongatus under conditions in which our proteins of interest or heme synthesis genes are overexpressed. A suitable candidate is the BG-11 medium modified by HK_SSC 2021 [7] with a higher manganese concentration.